The early visual system takes input, given to it in a language of photon catches and wavelength, and paraphrases it in a very different language of spike patterns that represent low-level properties of the visual scene. This information is combined and processed and recombined into progressively more complicated representations in a hierarchical system that ultimately provides the animal with what it needs to interact with the world. We know much about the response properties of individual neurons in visual cortex, about simple cells and complex cells and direction-selective cells and many others in a menagerie of richly varied circuit elements. But each of these cells is driven by a network of hundreds or thousands of other cells directly connected to it (a monosynaptic circuit), which provides it with all the information it will ever get about the outside world. We don’t know the functional properties of these circuits, the geometry and logic of their connectivity, or how single neurons take the spatiotemporal patterns of inputs bombarding their dendritic trees and transform them into outputs that feed into the next stage of processing. These are the basic questions of microcircuitry and single-neuron computation that inform my research, and by combining new circuit tracing methods with optical physiology in behaving animals, it may be possible to finally glimpse how these circuits are wired and function in vivo.